“It Works” Is a Trap

I’ve started to distrust the phrase “it works.” Not because it’s wrong—but because it usually hides a future problem. Most bugs I deal with aren’t from things that were broken. They come from code that worked fine… until someone had to change it. A quick example: function processUser(user) { if (user.type === "admin") { sendEmail(user.email); logActivity(user.id); } else if (user.type === "guest") { if (user.email) sendEmail(user.email); } } Nothing wrong here. You ship it. Then a month later: new user types show up email rules change logging becomes conditional Now this “simple” function is where changes go to die. What I’ve learned (the hard way): Every extra condition feels cheap at the time But it increases the cost of the next change And the next one And the next one Eventually, nobody wants to touch it. These days I try to optimize for changeability, not just correctness. Not overengineering—just small habits: clearer names fewer responsibilities per function making dependencies